A New Method for Identifying Traffic Accident Black Spots Avoiding Statistical Influence

In order to overcome the defects of the existing methods for identifying the road traffic accident black spots which were influenced by the statistical scale and the statistical starting and end points, this paper presents a new method based on the gray correlation degree which is not influenced by the statistical scale. First, the road is divided into some cells which are the same length, and then, the longest cell length is confirmed according to the proportion of the cells in which the accident number is just one. Secondly, that composing the reference sequence makes use of the longest seriate cells in which all the adjacent cells have accidents. And then, the seriate cells are taken as the dynamic comparison sequence whose length is the same as the reference sequence. Thirdly, the gray correlation degrees are calculated by using the accident numbers of the two sequences. Then, the threshold value for identifying the traffic accident black spots can be confirmed according to the accumulative frequency distribution figure of grey correlation degree. Finally, the location and severity of the traffic accident black spots are ascertained by using the figure of the mileage gray correlation degree. The results of the instance demonstrate that the method is feasible.